KR0183779B1 - Bill registering method and apparatus thereof - Google Patents

Abstract

Disclosed are a method capable of easily designating geometric information of a part to be recognized for any slip, contextual information of an object to be recognized, and an apparatus suitable therefor.

In the journal registration method according to the present invention, the journal registration method for inputting the geometric characteristics of the journal and the context information of the object to be recognized using computer graphics, the journal image input process of inputting an image of the journal to be registered; A document full property designation process for designating a property of the entire document, such as whether or not preprocessing is performed for the document image and rotation information; A region designation process of designating a characteristic of the region to be recognized in the document; And a document definition file output process of outputting a document definition file created according to a predetermined grammar of information registered in the document full property designation process and region designation process.

The journal registration method according to the present invention has the effect of easily and accurately performing a registration operation for any journal using computer graphics.

Description

Document registration method and device

1 is a flowchart showing a slip registration method according to the present invention.

FIG. 2 shows an example of a slip image input screen used in the slip image input process shown in FIG.

FIG. 3 shows an example of a dialog box used in the overall document property designation process shown in FIG.

4 shows an example of an area designation screen used in the area designation process shown in FIG.

5 is a diagram illustrating an example of a data area designation screen that appears when a data function is selected in the menu bar illustrated in FIG. 4.

FIG. 6 is a diagram illustrating an example of a mark region designation screen that appears when a mark function is selected in the menu bar shown in FIG.

7 is a block diagram showing the configuration of the slip registration apparatus according to the present invention.

The present invention relates to a method and apparatus for registering a journal, and more particularly, to a method capable of easily designating geometric information of a part to be recognized for a journal and context information of an object to be recognized, and an apparatus suitable therefor.

A voucher is a document that has a certain form, such as various bills or credit card receipts. The job of making a database to manage and keep them is called a slip entry.

At present, most of the document entry is performed by the worker directly inputting the contents of the document through the keyboard.

The manual document input method is expensive in terms of time and economics when the types of documents to be input are diversified and the quantity is large.

Automatic entry of a document can be divided into a document registration task for specifying a document format, a recognition task for recognizing the contents of the document using a specified document format, and a task for verifying the recognized results to make a database.

Among them, the registration of a slip is responsible for inducing the effective recognition by designating the geometric information of the part to be recognized and the context information of the object to be recognized.

Therefore, in order to enter various forms of documents quickly and accurately, an effective document registration method is required.

Known voucher registration methods up to now include a method of writing a description of a voucher to be input using a language having a certain syntax and a description using computer graphics.

The former method is adopted by an optical reading device for journal entry sold by NEC, Toshiba, etc. of Japan, and each time a new type of journal occurs, the coordinates of the recognition target area are measured using a measurer or the like. After measuring, rewrite it as a script file for a specific language.

The latter method is adopted by Nestor Inc.'s journal entry software product, which automatically converts user descriptions into script files through computer graphics.

Among the conventional methods, the method of writing the content describing the document to be input using a language with a certain rule has the trouble of re-measuring the geometric characteristics of the object to be recognized for each document. Errors can also occur.

In addition, the recognition contribution attribute that can be specified is extremely limited, so that whenever a new document entry is required, the recognizer should be modified to match the characteristics of the document.

The method adopted by Nestor et al. Eliminates some of the hassle caused by measuring the coordinates of the recognition target area, but does not completely cope with the document characteristics that are the core of the document registration, and uses attributes specific to a specific recognizer. Use is very limited.

An object of the present invention is to provide a method for easily designating geometric information of a part to be recognized for any slip and context information of an object to be recognized by using computer graphics technology. It is done.

Another object of the present invention is to provide a slip registration device suitable for the slip registration method described above.

In the journal registration method according to the present invention for achieving the above object, in the journal registration method for inputting the geometric characteristics of the journal and the context information of the object to be recognized using a computer graphic, the journal image for inputting an image of the journal to be registered Input process; A document full property designation process for designating a property of the entire document, such as whether or not preprocessing is performed for the document image and rotation information; Area specification process for specifying the characteristics of the recognition target area in the document; And a document definition file output process of outputting a document definition file created according to a predetermined grammar of information registered in the document full property designation process and region designation process.

According to another aspect of the present invention, a voucher registration apparatus includes: an input unit for inputting an image of a voucher to be registered; A document total property designation unit for designating rotation information, shift correction information, and form removal information for the entire document; A data area property designating unit that specifies attributes such as the number of characters of a recognition target area on a document, a character, a relationship with another data area, the number of detail areas, and preprocessing of data area units; A detail region property designation unit for specifying attributes such as a start position, an end position, a character type, a minimum / maximum value, and a dictionary / unigram name of the detail region; A mark designation unit for designating a mark area including a feature point which is always fixed with respect to the same kind of document in order to correct distortion generated when the document is input; And a document definition file generation unit for outputting a document definition file prepared according to a predetermined grammar of information registered in the document full property designation unit, the data area property designation unit, the detailed area property designation unit, and the mark designation unit. Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

1 is a flowchart showing a slip registration method according to the present invention. The apparatus shown in FIG. 1 includes a document image input process, a document overall property designation process, an area designation process, and a document definition file output process.

In the document image input process (step 100), an image of a document to be registered is input. The image of the slip may be provided through a scanner, an image processing device connected to the imaging device, or the like.

2 shows an example of a slip image input screen used in the slip image input process (step 100). In the document image input screen shown in FIG. 2, reference numeral 20 is a menu bar indicating a function item such as file input / output, scanner input, recognition / verification, utility, zoom in / out, rotation, file movement, and the like. This is the display area where the document is displayed.

The document full characteristic assignment process (step 120) specifies the characteristics of the entire document. The characteristics of the entire document can be specified through a dialog box as shown in FIG. 3, and there are preprocessing information and rotation information.

The presence or absence of preprocessing is a feature that takes into account when preprocessing of the entire document image is required, such as form removal, skew correction, and rotation correction, and the rotation information is entered as specifying the rotation state of the document. This is a characteristic considering the case where is printed sideways.

The area designation process (step 120) specifies the geometrical characteristics for the document registration. The geometrical characteristics include the characteristics of the recognition target region (hereinafter referred to as the data region) and the mark region characteristic.

The data area consists of one or more sub-areas, which also specify the characteristics for these sub-areas when the data area is specified.

The characteristics of the data area include the maximum number of characters, the character body, relationships with other data areas, and preprocessing operations.

The subregion characteristics include a start position and an end position indicating a character position in the data region, a type of character to be recognized, a minimum / maximum value, a name of a dictionary, and the like.

The mark area characteristic is used for correction of parallel movement distortion or rotation movement distortion of the document image to be input.

4 shows an example of an area designation screen used in the area designation process (step 120). In the slip image input screen shown in FIG. 4, reference numeral 40 is a menu bar indicating a function item such as a mark, data, or alignment.

5 is a diagram illustrating an example of a data area designation screen that appears when a data function is selected in the menu bar 40 illustrated in FIG. 4. Data area characteristics can be specified through a two-step dialog box as shown in FIG. 5 and include a maximum number of characters, a character, a relationship with other data areas, and a data processing unit.

The maximum number of characters means the maximum number of characters that the data area can have. The character font means a printing state of a character, and may be divided into, for example, a cursive (handwritten), printed (printed), mixed, and the like.

The relationship characteristic with other data areas may be used to correct or verify the recognition result of the recognizer by using an arithmetic relationship such as a sum or a product of other data areas.

Data area unit preprocessing includes items such as horizontal line removal, vertical line removal, isolation point removal, overlapping line removal, slope correction, and square removal, which are characteristics in case a preprocessing operation is required only for a specific data area.

In addition, a character separation method and a recognition method can be added as characteristics of the data area to select an appropriate method according to the data area.

Each data area can have one or more subregions (units), which can be divided into subdivisions according to subregion characteristics to maximize recognition efficiency. Character type (Hangul / Hanja / English (small) / English (large) / special character / number) that specifies the character to be recognized, minimum / maximum value that limits the range when the character type is a number, and pre-processor The dictionary / unigram name, etc., for the recognizer having the above are characteristic elements that constitute the detailed area.

6 is a diagram illustrating an example of a mark area designation screen that appears when a mark function is selected in the menu bar 40 shown in FIG.

The mark is used for the purpose of correcting various distortions such as the translational distortion and rotational movement distortion generated when the document is input through the scanner. The mark registration is used to mark a specific area of the document as shown in the example of FIG. Mark it as a rectangle and assign it to the mark area.

The mark area should be specified to include an independent object that always appears in a fixed position for homogeneous documents.

In the document registration method of the present invention, one feature point is found for the assigned mark area, and the modified coordinates are registered after the coordinates of the mark area are modified to be the center of the mark area.

When other documents of the same kind are entered, the registered mark area is analyzed to find the feature point, and the difference between the position of the feature point and the center point of the mark area can be used for various corrections.

Up to four mark areas on the document can be registered.

The document definition file output process 130 outputs the document definition file created according to a certain grammar of information registered in the document full property designation process 110 and the region designation process 120.

7 is a block diagram showing an embodiment of a slip registration apparatus according to the present invention. The apparatus shown in FIG. 7 is implemented on a computer. Accordingly, a monitor, a key input device, and the like are omitted since they are basic components of the computer.

The apparatus shown in FIG. 7 includes an input unit 70, a document full property designation unit 72, a data area property designation unit 74, a detailed area property designation unit 76, a mark designation unit 78, and a document definition. A file generation unit 80 is provided.

The input unit 70 inputs an image of a document to be registered and is implemented as a scanner.

The entire document property designation unit 72 performs preprocessing such as shift correction, form removal, and rotation correction for the entire document. Through the document full characteristic designation unit 72, the user instructs necessary correction processing while viewing an image input through the input unit 70 through a monitor (not shown).

The data area characteristic designation unit 74 designates attributes such as the number of characters in the recognition target area, characters, relationships with other data areas, the number of detail areas, and preprocessing in units of data areas.

The detail region property designation unit 76 designates attributes such as a start position, an end position, a character type, a minimum / maximum value, a dictionary / unigram name, and the like of the detail region.

The mark designation unit 78 designates a mark area including a feature point which is always fixed with respect to the same kind of document in order to correct distortion generated when the document is input.

In addition, the mark designation unit 78 searches for a feature point in the designated mark area, and inputs coordinates of the modified mark area so that the searched feature point is located at the center point of the mark area.

The document definition file generation unit 80 matches the information registered in the document full property designation unit 72, the data area property designation unit 74, the detailed area property designation unit 76, and the mark designation unit 78 according to a predetermined grammar. Output the document definition file you created.

Each component in the apparatus as shown in FIG. 7 may be composed of one module in a program executed on a computer.

As described above, the slip registration method according to the present invention has an effect that it is possible to easily and accurately register the slip on any slip using computer graphics.

In addition, by designating the recognition attribute to maximize the recognition efficiency, it is possible to maximize the performance of the recognizer, and has the advantage of being able to recognize any type of slip without modification of the recognizer.

Claims (7)

A document registration method for inputting geometric characteristics of a document and contextual information of a document to be recognized using computer graphics, comprising: a document image input process of inputting an image of a document to be registered; A document full property designation process for designating a property of the entire document, such as whether or not preprocessing is performed for the document image and rotation information; A region designation process of designating a characteristic of the region to be recognized in the document; And a document definition file outputting step of outputting a document definition file created according to a predetermined grammar of information registered in the document full property designation process and area designation process. The journal registration process according to claim 1, wherein the area designation process expresses data by attributes such as the number of characters in a recognition target area, a character body, a relationship with other data areas, the number of detailed areas, and preprocessing of data area units. Way. The method of claim 2, wherein the area designation process expresses data using attributes of a start position, an end position, a character type, a minimum / maximum value, a dictionary / unigram name, and the like of a detail region. The method of claim 1, wherein the area designation process further comprises a mark area designation process for designating a mark area including a feature point which is always fixed with respect to the same document in order to correct distortion generated when the document is input. Journal registration method characterized by the above. The method of claim 4, wherein the marking area designation process further comprises: a mark area modification process of searching for a feature point in the designated mark area and inputting coordinates of the modified mark area so that the found feature point is located at the center point of the mark area. Journal registration method characterized by the above. A journal registration device for inputting geometric characteristics of a journal and contextual information of a recognition object using computer graphics, comprising: an input unit for inputting an image of a journal to be registered; A document total property designation unit for designating rotation information, shift correction information, and form removal information for the entire document; A data area property designating unit that specifies attributes such as the number of characters of a recognition target area on a document, a character, a relationship with another data area, the number of detail areas, and preprocessing of data area units; A detail region property designation unit for specifying attributes such as a start position, an end position, a character type, a minimum / maximum value, and a dictionary / unigram name of the detail region; A mark designation unit for designating a mark area including a feature point which is always fixed with respect to the same kind of document in order to correct distortion generated when the document is input; And a document definition file generation unit for outputting a document definition file created according to a predetermined grammar of information registered in the document full property designation unit, the data area property designation unit, the detail area property designation unit, and the mark designation unit. The slip registration apparatus according to claim 6, wherein the mark designation unit searches for a feature point in a designated mark area and inputs coordinates of the mark area modified so that the searched feature point is located at the center point of the mark area.